Assessment and pharmacological correction of abnormalities in chloride (Cl-), bicarbonate (HCO3-) and mucus transport in rectal biopsies of CF patients

Accumulation of viscid mucus in the lung, intestine, pancreas, hepatobiliary
tract, and reproductive tract is a hallmark of cystic fibrosis (CF;
*mucoviscidosis*) and is the primary cause of defective mucociliary clearance
in the airways and of luminal obstruction in the GI and reproductive tract.
Recent studies in CF mouse models indicate that normal mucus release in
intestinal epithelium requires concurrent bicarbonate (HCO3-) secretion and
that aggregated mucus is a consequence of defective transepithelial HCO3-
transport in CF [1]. Moreover the pancreatic phenotype of CF patients
segregates well with mutations in the gene for CF transmembrane conductance
regulator (CFTR) that severely disrupt CFTR-dependent HCO3- transport [2].
However, in contrast to the wealth of information about the chloride transport
defect in native epithelia from both CF mice and CF patients emerging from in
vivo and ex vivo assays (e.g. sweat test; measurements of nasal potential
difference, NPD; intestinal current measurements in human rectal biopsies, ICM
[3-9]), there is as yet a paucity of data about bicarbonate secretion and mucus
release in native epithelia from both healthy individuals and CF patients.
Whereas sweat tests and NPD measurements are only suitable to monitor possible
defects in Cl- and Na+ transport, ICM performed in mini-Ussing chambers is the
sole technique capable of measuring both electrogenic Cl- and HCO3- secretion.
However sofar this technique has been developed and exploited by us [3-6] and
others[7,8] solely for diagnostic purposes, and a possible contribution of
HCO3- to the transepithelial anion current has not been evaluated yet or was
even precluded by the use of HCO3- -free bath fluid in the Freiburg protocol
[7,8].
Traditionally, intestinal bicarbonate secretion has been studied mainly or
exclusively in the duodenum (DBS) [9,10], ignoring potential contributions from
the colon. However, very recent studies in distal colon of CF mice and of mice
lacking bestrophin-2 (Best2), a Ca2+-activated anion channel, have shown that
(1) a large part of the plateau phase of carbachol-induced anion currents is
HCO3- -and Best2-dependent, and (2) Best2 is localized basolaterally in mouse
and human colonic goblet cells and, in concert with apical Cl-/HCO3-
exchangers, may enable Ca2+-stimulated transepithelial HCO3- secretion in
parallel with mucin secretion [11] . In this way, HCO3- may promote mucin
volume expansion and release by raising the local pH and sequestering Ca2+
[12]. In addition, colonic crypt cells may secrete HCO3- through basolateral
NBC1 Na+ / HCO3- cotransporters and apical CFTR channels, and brushborder cells
at the epithelial surface may cause net secretion of HCO3- by the apical
SLC26a3 Cl-/HCO3- exchanger under conditions that the NHE3 Na+/H+ exchanger is
inhibited by Ca2+-, cAMP-or cGMP-linked secretagogues [10,13].

Importantly, colonic HCO3- secretion is expected to be reduced or abolished in
CF at multiple levels: ill-functioning of CFTR, acting itself as a HCO3-
channel or as a Cl- shunt pathway to facilitate SLC26-mediated HCO3-
secretion10 ; reduced expression and redistribution of Best2[11] ; and reduced
expression of HCO3- -producing enzymes, i.e. the carbonic anhydrases (Car4,
Car2) in mouse colon [14].

Sofar, studies of colonic HCO3- secretion and mucus release have been carried
out exclusively in rodent models, and data on human colon are completely
lacking. Furthermore, colonic HCO3- secretion has been assessed only by
transepithelial current measurements, ignoring the contribution of
electroneutral components.
Surprisingly, even less information is as yet available about the ability of
the novel CFTR correctors emerging from high-throughput screens (e.g. the
Vertex and Verkman correctors, including VX-809 that is presently tested in
clinical trials) to correct defects in Cl-, HCO3- and mucin release in
addition to their known ability to improve transepithelial Cl- secretion [15.]
Unfortunately none of the outcome parameters monitored in the clinical trials
sofar (NPD, sweat test, FEV1, etc.) provides direct information on HCO3- and
mucin secretion in the CF affected epithelia. Therefore, additional studies
focusing on such measurements in an easily accessible tissue, i.e. rectal
biopsies, are urgently needed.

We aim to fill up the present gap in knowledge about HCO3- and mucus transport
in native human intestine, and about possible effects of pharmacological
correctors on these parameters in CF, by detailed electrophysiological and
biochemical ex vivo studies in rectal biopsies from CF patients and healthy
subjects mounted in Ussing chambers
[3-6].
The following questions will be investigated:

1. How much of the basal and secretagogue-stimulated anion current in non-CF
and CF biopsies is carried by Cl- and how much by HCO3-?
Comparative current measurements in the presence and absence of HCO3- and
carbonic anhydrase (Car) inhibitors in the serosal bath will provide insight
into the electrogenic component of HCO3- secretion in CF and non-CF biopsies.
It is anticipated that electrogenic HCO3- secretion, in analogy to mouse distal
colon, is reduced in CF, by multiple causes: defective transport through the
CFTR channel itself; defective basolateral transport through the Best2 anion
channel in goblet cells which is shown to be downregulated and redistributed in
CF[11]; and downregulation of carbonic anhydrases in CF as a consequence of
pathological reprogramming of gene expression caused by defective stimulation
of the transcription factor PPAR-γ [14].

2. Does an electroneutral component contribute to colonic HCO3- secretion and
is this component altered in CF and modulated by colonic secretagogues?
The electroneutral component (most plausibly reflecting SLC26a3-mediated
Cl-/HCO3- exchange unbalanced by NHE3-mediated Na+/H+ exchange) can be infered
from pH-stat measurements of total HCO3- secretion (electrogenic+
electroneutral) into the luminal bath followed by subtraction of the
electrogenic component. We expect that this component is enhanced by
secretagogues capable of inhibiting NHE3 but not SLC26a3, i.e. Ca2+-, cAMP- and
cGMP-linked hormones, neurotransmitters and microbial enterotoxins [10,13] .
Whether this component is reduced too in CF rectal biopsies is more difficult
to predict considering the controversial evidence for a loss of cAMP-inhibition
of NHE3 in CF intestine (reported to occur in the jejunum of CF mice but not in
the jejunum of CF patients [16]). In addition, the effect of specific
pharmacological inhibitors of NHE3 on colonic HCO3- and mucin secretion will be
examined, encouraged by recent studies in Cftr-/- mice showing that the
additional KO of NHE3 restored luminal hydration, mucin release and (most
plausibly) HCO3- secretion and protected the mice against lethal intestinal
obstruction [17].

3. Is the calcium-activated anion channel bestrophin-2 involved in HCO3-
secretion in human colon, and down-regulated/redistributed in distal colon of
CF patients?
In human distal colon, similar to mouse colon, Best2 is localized in the
basolateral membrane of the mucin-secreting goblet cells[14], but it is unclear
whether human Best2 serves as a HCO3- importer and is likewise downregulated by
~80% in CF colon, as reported for Cftr-/- mice[14]. Because the survival time
of rectal biopsies (~24h) is too short to allow efficient siRNA-induced
knockdown (KD) of Best2, the use of pharmacological inhibitors is the only way
to evaluate the contribution of Best2 to HCO3- secretion in human colon.
Surprisingly, the cyclo-oxygenase inhibitor indomethacin was recently shown to
act as a rather specific inhibitor of Best2 anion channels[14], allowing us to
assess the contribution of Best2 to carbachol/Ca2+-induced secretion under
conditions in which we compensate for the simultaneous indomethacin-induced
inhibition of cyclo-oxygenases/prostaglandin synthesis by adding PGE2
exogenously. This compensation is needed because endogenous prostaglandins in
human rectal biopsies are supposed to partially activate CFTR through cAMP
signaling and to allow carbachol to stimulate Cl- secretion through the CFTR
channel indirectly by Ca2+-stimulation of basolateral K+ channels that
hyperpolarize the membrane and increase the driving force for apical Cl- exit
[11,13]. Moreover confocal microscopy will be applied to determine possible
changes in expression and localization of Best2 in the CF biopsies.

4. Can the defect in HCO3- secretion and mucin release in CF colon be corrected
by pharmacological CFTR correctors?
Our recent success in preserving the morphology and function of human rectal
biopsies for at least 24 h ex vivo will allow us to answer the important
question whether pharmacological CFTR correctors that emerged from
high-throughput screening (HTS) and are known to partially restore
CFTR-mediated Cl- secretion in CF epithelial cells are also capable of
restoring CF defects in epithelial HCO3- and mucin secretion. These
experiments will reveal whether rescued F508del-CFTR in the apical membrane is
able to conduct HCO3- itself or can normalize CFTR-dependent, SLC26A-mediated
HCO3- transport.

5. Can the defect in HCO3- secretion and mucin release in CF human colon be
corrected by PPAR-γ agonists in the absence of CFTR correction?
Recent studies in Cftr-/- mice have shown that the defect in heat-stable
enterotoxin (STa; a cGMP-linked secretagogue)-induced colonic HCO3- and mucin
secretion could be restored, and intestinal obstruction prevented, by in vivo
treatment of the mice with the PPAR-* agonist rosiglitazone[14]. This
correction was associated with the upregulation of the carbonic anhydrases Car4
and Car2, encoded by genes that are regulated by the PPAR-γ signaling
pathway[14]. An important aim of this project is to verify whether PPAR-γ
agonists are able to exert a similar action on human distal colon ex vivo.
Aside rosiglitazone, we will also test the bicyclic fatty acid lubiprostone[18]
(presently in clinical use against chronic obstipation) for its ability to
cross-activate PPAR-* and to restore HCO3- and mucin secretion in CF biopsies.

1. Studying HCO3- secretion, mucus stasis and the expression and localization
of key transporters and enzymes involved in HCO3- secretion in rectal biopsies
of healthy subjects and CF patients.

Transepithelial current measurements in rectal suction biopsies from healthy
individuals and CF patients (4 per patient) will be performed in the
*Rotterdam* mini-Ussing chambers (aperture 1.13 mm2) connected to a DVC-1000
voltage clamp (WPI) and a Powerlab for digitalization and current recording
(LabChart software), as described in detail in ref. 6. The serosal bath
solution (Meyler buffer containing 10 mM glucose) is circulated with a carbogen
gas-lift system (95% O2, 5% CO2) at pH 7.4, and the unbuffered solution at the
luminal side (154 mM NaCl containing 10 µM amiloride to inhibit ENaC channels)
is circulated with 100% O2. Following a 1h equilibration period, the
electrogenic component of basal and secretagogue-stimulated anion (i.e. HCO3- +
Cl-) transport will be assessed by short-circuit current (Isc) measurements,
and total HCO3- secretion (i.e. electrogenic+electroneutral) is monitored by
measuring luminal alkalinization using a continuous pH-stat titration method
(Radiometer, Copenhagen). By this approach the luminal pH is maintained at 7.4
by addition of an isotonic solution containing 2 mM HCl. Subsequently, the
secretagogues are washed out by multiple changes of the bath fluid, and the Isc
measurements are repeated after isotonic replacement of HCO3- serosally by
Na-HEPES (pH 7.4) and Na-gluconate, bilateral addition of the Car inhibitor
methazolamide (100 µM), and a shift to 100% O2 gassing. The electrogenic
component of HCO3- transport is assessed by comparing the Isc responses in the
first and second period, i.e. in the presence and absence of serosal HCO3-.
To evaluate the HCO3- secretory response to different intracellular signaling
pathways, the following secretagogues will be tested consecutively: carbachol
(Ca2+-linked)-wash out-guanylin (cGMP-linked)-STa (cGMP-linked)-wash out-
forskolin+ IBMX (cAMP-linked)-genistein (improving the gating of
F508del-CFTR)-carbachol.

In separate experiments, the contribution of the Ca2+-dependent anion channel
bestrophin-2 to HCO3- secretion will be estimated by comparing Isc and pH-stat
responses to carbachol in the absence or presence of the Best2 inhibitor
indomethacin [11]. PGE2 will be added to mask indomethacin-inhibition of
endogenous PGE formation.

Mucus stasis will be assessed by Alcian Blue staining of the biopsies, and
expression and (re)distribution of transporters (CFTR, Best2) and enzymes
(Car4, Car2) will be examined by immunostaining and confocal microscopy.

We estimate that completion of part 1 of this study will require the
participation of 10 healthy controls and 8 F508del CF patients (5
biopsies/patient).

2. Ex vivo repair of CF defects in colonic HCO3- secretion and mucus release by
pharmacological approaches.

We have recently defined tissue preservation conditions in which human rectal
biopsies mounted in Ussing chamber inserts and maintained at 370C under
carbogen gassing remain fully functional for at least 24h. We aim to exploit
these findings in this project by testing long-term rescue effects of 3 classes
of potential pharmacological correctors on rectal biopsies from homozygous
F508del CF patients ex vivo:

(i) Correctors of F508del misprocessing that are known to improve
CFTR-mediated Cl- transport but have not been tested yet for their ability to
improve HCO3- and mucus transport in CF. The Vertex corrector VRT-809,
previously found to enhance CFTR-mediated Cl- secretion in human bronchial
epithelial cells [15], and presently evaluated in clinical trials, will be
tested first as a positive control. Subsequently, 4 recently developed EPIX
F508del-CFTR correctors (donated by the CFFT-USA) will be investigated in the
same test protocol. Following the Isc and pH-stat measurements in fresh
biopsies described under (1), biopsies will be incubated for 24h in the
presence of the CFTR corrector or vehicle (2 biopsies/condition), followed by
repetition of the Isc and pH-stat measurements and Alcian Blue staining.

(ii) Potential correctors of the defect in PPAR-gamma signaling in CF, i.e.
the synthetic PPAR-* ligand rosiglitazone [14] and the bicyclic fatty acid and
PGE1 derivative lubiprostone [18], both in clinical use for treatment of type-2
diabetes and chronic obstipation, respectively.

(iii) S3226 (provided by Sanofi Aventis), a novel inhibitor of NHE3 that may
mimic the beneficial effects of NHE3-KO on CF pathology (e.g. improvement of
intestinal HCO3- secretion) as shown previously in Cftr-/- mice [17] (see ii
for test protocol).

We estimate that completion of part 2 of this study (involving the testing of 8
different pharmacological correctors) will require the participation of 32
F508del CF patients. Power analysis shows that the demonstration of a
statistically significant improvement of transepithelial anion secretion in
the Ussing chamber (up to a Isc value >10% of healthy controls) requires the
testing of each corrector or vehicle in 2 biopsies from at least 4 CF
patients.

Recruitment of healthy controls and CF patients:
Following written informed consent and with approval from the Medical Ethical
Committee of the UMCU, rectal biopsies (5/subject) from adult homozygous
F508del CF patients and healthy volunteers will be collected by a suction
biopsy device (Meeker Instruments, Utrecht) and processed exactly as described
previously. [18]

Burden and risk of rectal biopsy are minimal.

Risk of bleeding due to rectal biopsy in literature:
N= 389, comp. 2 (bleeding 0.5%) [19]